Near2009chap45.Pdf

Notothenioid fi shes (Notothenioidei)

Thomas J. Near and A lled vacant niches aJ er the onset of polar condi- Department of Ecology and Evolutionary Biology & Peabody tions ~35 Ma (2). 7 e fossil A shes preserved in the Eocene Museum of Natural History, Yale University, New Haven, CT 06520, La Meseta Formation on Seymour Island at the tip of the USA ([email protected]) Antarctic Peninsula indicate that before the development of polar conditions the nearshore A sh fauna of Antarctica Abstract was diverse, cosmopolitan, and not dominated by noto- thenioids (5). 7 e only documented notothenioid fossil Notothenioids are a clade of acanthomorph teleosts that is a well-preserved neurocranium of the extinct species represent a rare example of adaptive radiation among mar- Proeleginops grandeastmanorum from the La Meseta ine fi shes. The notothenioid Antarctic Clade is character- Formation that is dated to ~40 Ma (6–10). ized by extensive morphological and ecological variation Ecologically, Antarctic notothenioids have diversiA ed and adaptations to avoid freezing in the ice-laden water of into both benthic and water column habitats (2). Several Southern Ocean marine habitats. A recent analysis of noto- lineages are able to utilize water column habitats des- thenioid divergence times indicates that the clade dates to pite lacking a swim bladder by modiA cation of buoyancy the Cretaceous (125 million years ago, Ma), but the Antarctic through the reduction of ossiA cation and the evolution of Clade diversifi ed near the Oligocene–Miocene boundary intra- and intermuscular lipid deposits (11, 12). A notable (23 Ma). These age estimates are consistent with paleogeo- group of notothenioid species is the Channichthyidae, graphic events in the Southern Ocean that drove climate or iceA shes (Fig. 1). Species in this clade are also called change from temperate to the polar conditions observed the “white-blooded” A shes, because of the absence of the today. oxygen-transporting molecule hemoglobin, which is the result of deleted globin loci possibly initiated by inter- Notothenioids represent an adaptive radiation of teleost speciA c hybridization and subsequent introgression (13). A shes in the frigid waters of the Southern Ocean sur- 7 ese are the only vertebrates that exhibit this bizarre rounding Antarctica (1). Of the ~129 recognized species, phenotype and it is thought that the persistence of this 101 are found in marine costal habitats of Antarctica apparently deleterious trait is due to the cold oxygen-sat- (Fig. 1), and the remaining species are distributed along urated water that provides adequate oxygen via passive costal areas of southern South America, the Falkland diB usion into the body (14). Islands, southern New Zealand, southern Australia, and Tasmania (2). In addition to a diverse array of adapta- tions to survive in the freezing Antarctic marine habi- tats, notothenioids are unique in that they completely dominate the A sh fauna of the Southern Ocean. Among benthic A sh samples taken on the Antarctic shelf, noto- thenioids comprise nearly 77% of the species diversity, more than 91% of the species abundance, and ~91% of the biomass (3). A hypothesized key innovation that facilitated the diversiA cation of Antarctic notothenioids is the origin of an antifreeze glycoprotein from a tyripsinogen-like ancestral gene that confers protection from freezing in the subzero Southern Ocean waters (4). 7 e ecological Fig. 1 A 29.2 cm long (standard length) channichthyid and morphological diversity of Antarctic notothenioids notothenioid (Chionodraco myersi: YPM 16533) sampled from is extensive and it is thought that the clade diversiA ed the Bransfi eld Strait. Credit: T. J. Near.

T. J. Near. Notothenioid A shes (Notothenioidei). Pp. 339–343 in e Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

HHedges.indbedges.indb 339339 11/28/2009/28/2009 11:28:13:28:13 PPMM 340 THE TIMETREE OF LIFE

Channichthyidae 8 Bathydraconidae-2 7 Bathydraconidae-1 6 Artedidraconidae 9 5 Harpagiferidae 4 Nototheniidae-3 Nototheniidae-2 3 2 Nototheniidae-1 1 Pseudaphritidae Bovichtidae

Early K Late KPaleogene Neogene MESOZOIC CENOZOIC 125 100 75 50 25 0 Million years ago

Fig. 2 A timetree of notothenioid fi shes (Notothenioidei). Divergence times are shown in Table 1. Abbreviation: K (Cretaceous).

7 e monophyly of Notothenioidei has been supported radiation at 57–45 Ma (28). 7 e range of estimated diver- in phylogenetic analyses of morphological characters gence times for the common ancestor of the Antarctic (15, 16) and DNA sequences from mitochondrial and Clade was 21–5 Ma (4, 21, 26–28, 35). Several of these nuclear genes (17–20). Phylogenetic relationships among studies also estimated the divergence times of particu- lineages within Notothenioidei inferred from analyses lar lineages within the Antarctic Clade. Age estimates for of morphological and molecular data sets are generally the Channichthyidae ranged between 23.3 and 2 Ma (26, consistent with traditional taxonomic hypotheses devel- 31, 36), and molecular age estimates for the nototheniid oped during the time of great Antarctic exploration in clade Trematomus (with and without Trematomus scotti) the early twentieth century. Taxonomically, eight fam- were 3.4–2.8 Ma (26, 34). ilies are recognized in the Notothenioidei and all but the Any eB ort that attempts to estimate divergence times Bathydraconidae (DragonA shes) and Nototheniidae were using molecular data will face a specter of uncertainty resolved as monophyletic groups in molecular phylogen- and ever present confounding variables, and all of these etic analyses (21–28). Monophyly of Nototheniidae was notothenioid molecular divergence time studies suf- supported in morphological phylogenetic analyses ( 29), fer minimally from one of three severe methodological and phylogenetic analyses of complete mtDNA 16S rRNA problems. First, calibration is a major concern as most (23). Other phylogenetic analyses have focused on speciA c estimates of notothenioid divergence times rely on exter- notothenioid subclades, including the Channichthyidae nal or “universal” rates of DNA sequence evolution esti- (30–33), Bathydraconidae (22), Artedidraconidae (15), mated for other clades and applied to notothenioids (4, and Nototheniidae (24, 34). One important result 27, 31, 34–37). 7 e extreme example of this strategy was from these phylogenetic investigations is the consist- the application of the rate of mtDNA evolution in trout ent monophyly of the Antarctic Clade (Nototheniidae, and salmon to estimate the age of the Antarctic Clade Harpagiferidae, Artedidraconidae, Bathydraconidae, from pairwise genetic distances calculated from a nucle- and Channichthyidae) that comprises the major lineages ar-encoded intron (4). 7 is strategy is problematic and of notothenioids that are found in the Southern Ocean ill-advised, because animal mtDNA genes have a much south of the Antarctic Polar Front (25, 26). higher nucleotide substitution rate than any sampled Eleven published studies have presented molecular nuclear gene, including introns (38). divergence time estimates for notothenioid clades. 7 e Even when paleogeographic calibrations have been ages estimated from these molecular analyses were quite used, they do not represent current hypotheses in the broad, but fairly similar across studies. One study esti- geological literature. For example, the separation of New mated the divergence time of the entire notothenioid Zealand from East Gondwana is given as 57 Ma in ref.

HHedges.indbedges.indb 340340 11/28/2009/28/2009 11:28:15:28:15 PPMM Eukaryota; Metazoa; Vertebrata; Actinopterygii; Teleostei; Notothenioidea 341

Table 1. Divergence times (Ma) and their credibility/confi dence intervals (CI) among notothenioid fi shes (Notothenioidei).

Timetree Estimates Node Time Ref. (4) Ref. (21) Ref. (26) Ref. (27) Ref. (28) Ref. (35) Ref. (42) Time Time Time Time Time Time Time CI

1125.0––––57–45–125.0129–121 247.0––––––47.048.4–45.6 3 24.1 14–5 12–8 21 16–10 15–11 15–7 24.1 24.6–23.6 422.4––––––22.422.9–21.9 518.9––––––18.919.3–18.5 617.0––––––17.017.4–16.6 716.1–35–15––––16.116.4–15.8 815.8––––––15.816.1–15.5 9 12.7 – – – – – – 12.7 –

Note: Node times in the timetree are from ref. (42).

(28) and used to calibrate the divergence of Bovichtidae and the Antarctic Clade. Penalized likelihood was used from other notothenioids. However, it appears that to correct for molecular evolutionary rate heterogeneity 80 Ma is a more appropriate age for this event (39, 40). among lineages and conA dence intervals were calculated In another study, the fragmentation of Australia and with a bootstrapping method. 7 e time-calibrated phyl- Antarctica is presented as occurring 38 Ma and used to ogeny is presented in Fig. 2 and the divergence times are calibrate the divergence of Pseudaphritidae and all other shown in Table 1. 7 e clade Nototheniidae is not mono- notothenioids (26). However, a range of 52–35 Ma is the phyletic in t his ana lysis, and t he timetree in Fig. 2 contains more appropriate age for this paleogeographic event (39, three nototheniid clades: clade 1 containing all the spe- 40). Second, most molecular estimates of divergence cies sampled from Dissostichus, Notothenia, Aethotaxis, times in notothenioids have ignored heterogeneity of Lepidonotothen, Patagonotothen, Trematomus, and molecular evolutionary rates among lineages. However, Indonotothen; clade 2 containing Gobionotothen gib- a few studies have used relative rate tests, where spe- berifrons and Gobionotothen acuta; and clade 3 con- cies exhibiting departure from rate heterogeneity were taining Pleuragramma antarcticum. Bathydraconidae excluded from the analysis (26–28). Relative rate tests are is also not monophyletic and the species are distributed problematic, because they measure the substitution rate between two clades: clade 1 contains Gymnodraco acu- in only a small portion of the phylogeny and the statis- ticeps, and clade 2 contains Cygnodraco mawsoni and tical signiA cance of relative rate tests must be corrected Parachaenichthys charcoti. when multiple tests are used (41). 7 ird, divergence time Most of the estimated divergence times from this estimates have oJ en been presented without conA dence penalized likelihood rate smoothed molecular phyl- or credibility intervals. ogeny are older than age estimates resulting from ana- 7 e collective problems exhibited among these lyses of pairwise genetic distances. For example, the notothenioid divergence time estimates were directly molecular divergence time estimate for the common addressed in a study that used a fossil calibration and a ancestor of Notothenioidei is ~125 Ma, more than double tree-based method to account for rate heterogeneity (42). the single previous molecular estimate (28). One study A molecular phylogeny and branch lengths were esti- used the fragmentation of Antarctica and Australia as mated from mtDNA gene sequences sampled from the a calibration set at 38 Ma for the common ancestor of 12S and 16S rRNA genes using maximum likelihood. 7 e Pseudaphritidae and the remaining notothenioids (26); fossil P. grandeastmanorum was used to provide a A xed however, the penalized likelihood estimated age for this minimal age estimate of 40 Ma for the node that repre- node is substantially older (Fig. 2; Table 1). Perhaps of sents the most recent common ancestor of Eleginopidae greatest interest to comparative biologists is the age of

HHedges.indbedges.indb 341341 11/28/2009/28/2009 11:28:15:28:15 PPMM 342 THE TIMETREE OF LIFE

the common ancestor of the Antarctic Clade, since this is 5. J. T. Eastman, Antarctic Sci. 12, 276 (2000). the lineage that exhibits adaptations to polar conditions 6. J. A. Case, in Geology and Paleontology of Seymour Island, such as the presence of antifreeze glycoproteins. Previous Antarctic Peninsula, R. M. Feldman, M. O. Woodburne, estimates using pairwise genetic distances had esti- Eds. (Geological Society of America, Boulder, 1988), pp. mated this clade diversiA ed between 21 and 5 Ma (4, 35) 523–530. 7. J. A. Case, M. O. Woodburne, D. S. Chaney, in Geology (Table 1). 7 e penalized likelihood estimate for the age and Paleontology of Seymour Island, Antarctic Peninsula, of this clade is older and close to the Oligocene–Miocene R. M. Feldman, M. O. Woodburne, Eds. (Geological boundary at 23 Ma (Fig. 2; Table 1). 7 is indicates that Society of America, Boulder, 1988), pp. 505–521. the Antarctic Clade diversiA ed aJ er the development 8. M. O. Woodburne, W. J. Zinsmeister, J. Paleontol. 58, 913 of the unrestricted Antarctic Circumpolar Current and (1984). Antarctic Polar Front, which formed aJ er the open- 9. J. T. Eastman, L. Grande, Antarctic Sci. 3, 87 (1991). ing of the Drake Passage between South America and 10. A. V. Balushkin, Vopr. Ikhtiol. 34, 298 (1994). Antarctica in the late Eocene (43). 11. A. L. DeVries, J. T. Eastman, Nature 271, 352 (1978). 7 e available divergence time estimates in notothen- 12. J. T. Eastman, A. L. DeVries, Copeia 1982, 385 (1982). ioids are far from providing the last word on the topic. 13. T. J. Near, S. K. Parker, H. W. Detrich, Mol. Biol. Evol. 23, Molecular age estimates for notothenioids need explor- 2008 (2006). ation with data sets that are much larger in terms of the 14. G. di Prisco, E. Cocca, S. K. Parker, H. W. Detrich, Gene 295, 185 (2002). number of genes and species sampled, relative to the par- 15. R. R. Eakin, in Antarctic Research Series, Vol. 31, Biology tial mtDNA gene sequences used in all previous studies. of the Antarctic Seas IX, L. S. Kornicker, Ed. (American In addition, the consistency of the P. grandeastmanorum Geophysical Union, Washington, 1981), pp. 81–147. fossil calibration should be assessed in a cross-validation 16. P. A. Hastings, in A History and Atlas of the Fishes of the analysis with external fossil calibrations sampled from Antarctic Ocean, R. G. Miller, Ed. (Foresta Institute for other acanthomorph teleost clades. 7 ese future investi- Ocean and Mountain Studies, Carson City, 1993), pp. gations also need to utilize strategies of divergence time 99–107. estimation that account for heterogeneity of molecular 17. W.-J. Chen, C. Bonillo, G. Lecointre, Mol. Phylogenet. evolutionary rates, as well as uncertainty in the fossil Evol. 26, 262 (2003). calibrations. Such analyses will provide reliable credibil- 18. A. Dettaï, G. Lecointre, Antarctic Sci. 16, 71 (2004). ity intervals for the molecular age estimates. 7 e non- 19. A. Dettaï, G. Lecointre, C. R. Biol. 328, 647 (2005). 20. W. L. Smith, M. T. Craig, Copeia 2007, 35 (February 28, parametric bootstrap procedure used in the penalized 2007). likelihood estimate of notothenioid divergence times 21. L. Bargelloni, L. Zane, N. Derome, G. Lecointre, does not account for uncertainties in the fossil calibra- T. Patarnello, Antarctic Sci. 12, 259 (2000). tions and likely results in misleadingly narrow conA dence 22. N. Derome, W.-J. Chen, A. Dettai, C. Bonillo, G. Lecointre, intervals (44, 45). Robust divergence time estimates for Mol. Phylogenet. Evol. 24, 139 (2002). notothenioids will facilitate investigations of the role of 23. T. J. Near, J. J. Pesavento, C. H. C. Cheng, Mol. Phylogenet. climate change and adaptive evolution in the diversiA ca- Evol. 32, 881 (2004). tion of this Antarctic adaptive radiation. 24. S. Sanchez, A. Dettai, C. Bonillo, C. Ozouf-Costaz, H. W. Detrich, G. Lecointre, Polar Biol. 30, 155 (2007). 25. T. J. Near, C. H. C. Cheng, Mol. Phylogenet. Evol. 47, 832 Acknowledgment (2008). 26. L. Bargelloni, S. Marcato, L. Zane, T. Patarnello, Syst. T.J.N. is supported by the U.S. National Science Biol. 49, 114 (2000). Foundation. 27. L. Bargelloni, G. Lecointre, in Fishes of Antarctica: A Biological Overview, G. D. Prisco, E. Pisano, A. Clarke, Eds. (Springer-Verlag, Berlin-Heidelberg, 1998), pp. References 259–273. 1. A. Clarke, I. A. Johnston, Trends Ecol. Evol. 11, 212 (1996). 28. L. Bargelloni, T. Patarnello, P. A. Ritchie, B. Battaglia, 2. J. T. Eastman, Antarctic Fish Biology: Evolution in a A. Meyer, in Antarctic Communities, B. Battaglia, Unique Environment (Academic Press, San Diego, 1993). J. Valencia, D. W. H. Walton, Eds. (Cambridge University 3. J. T. Eastman, Polar Biol. 28, 93 (2005). Press, Cambridge, 1997), pp. 45–50. 4. L. Chen, A. L. DeVries, C.-H. Cheng, Proc. Natl Acad. 29. A. V. Balushkin, J. Ichthyol. 40, S74 (2000). Sci. U.S.A. 94, 3811 (1997). 30. T. Iwami, Mem. Nat. Inst. Polar Res. Tokyo 36, 1 (1985).

HHedges.indbedges.indb 342342 11/28/2009/28/2009 11:28:15:28:15 PPMM Eukaryota; Metazoa; Vertebrata; Actinopterygii; Teleostei; Notothenioidea 343

31. W.-J. Chen, C. Bonillo, G. Lecointre, in Fishes of 3 8 . W . - H . L i , Molecular Evolution (Sinauer, Sunderland, Antarctica: A Biological Overview, G. Di Prisco, E. Pisano, Massachusetts, 1997). A. Clarke, Eds. (Springer-Verlag, Berlin-Heidelberg, 39. I. Sanmartín, F. Ronquist, Syst. Biol. 53, 216 (2004). 1998), pp. 287–298. 40. A. G. Smith, D. G. Smith, B. M. Funnell,Atlas of Mesozoic 32. T. J. Near, J. J. Pesavento, C.-H. C. Cheng, Mol. Phylogenet. and Cenozoic coastlines (Cambridge University Press, Evol. 28, 87 (2003). Cambridge, 2004). 33. O. S. Voskoboinikova, Zool. Zh. 79, 321 (2000). 41. M. J. Sanderson, in Molecular Systematics of Plants, II: 34. P. A. Ritchie, L. Bargelloni, A. Meyer, J. A. Taylor, J. A. DNA Sequencing, D. E. Soltis, P. S. Soltis, J. J. Doyle, Eds. Macdonald, D. M. Lambert, Mol. Phylogenet. Evol. 5, 383 (Kluwer Academic Publishers, Amsterdam, 1998), pp. (1996). 242–264. 35. L. Bargelloni, P. A. Ritchie, T. Patarnello, B. Battaglia, 42. T. J. Near, Antarctic Sci. 16, 37 (2004). D. M. Lambert, A. Meyer, Mol. Biol. Evol. 11, 854 43. H. D. Scher, E. E. Martin, Science 312, 428 (2006). (1994). 4 4 . J . L . 7 orne, H. Kishino, in Statistical Methods in 36. J. T. Eastman, A. R. McCune, J. Fish. Biol. 57, 84 Molecular Evolution, R. Nielsen, Ed. (Springer, New (2000). York, 2005), pp. 233–256. 37. A. Stankovic, K. Spalik, E. Kamler, P. Borsuk, 45. Z. Yang, Computational Molecular Evolution (Oxford P. Weglenski, Polar Biol. 25, 203 (2002). University Press, Oxford, 2006).

HHedges.indbedges.indb 343343 11/28/2009/28/2009 11:28:16:28:16 PPMM